Research Article
Using Principal Component Analysis to Solve a Class Imbalance Problem in Traffic Incident Detection
Table 3
Performance comparison between PCC and PLSR.
| Classifiers | DR | FAR | MTTD | CR | Significance | Explained | Bound |
| 0.7 | 70% | 0.2 | 75.76 | 10.78 | 18.08 | 87.16 | 0.3 | 69.70 | 10.75 | 12.65 | 87.17 | 90% | 0.2 | 69.70 | 10.40 | 18.57 | 87.53 | 0.3 | 63.64 | 10.31 | 12.29 | 87.58 |
| 0.8 | 70% | 0.2 | 81.82 | 12.48 | 15.63 | 85.59 | 0.3 | 72.73 | 12.04 | 12.21 | 85.97 | 90% | 0.2 | 72.73 | 12.10 | 16.29 | 85.93 | 0.3 | 63.64 | 11.65 | 11.76 | 86.32 |
| 0.9 | 70% | 0.2 | 81.82 | 14.37 | 11.70 | 83.81 | 0.3 | 72.73 | 13.42 | 11.92 | 84.68 | 90% | 0.2 | 75.76 | 13.93 | 11.64 | 84.22 | 0.3 | 66.67 | 13.01 | 11.18 | 85.05 |
| Average of PCC | 75.76 | 13.93 | 11.64 | 84.22 | PLSR | 69.70 | 13.07 | 11.04 | 85.12 |
|
|